| Message ID | 1436549771-6097-1-git-send-email-sudeep.holla@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 10/07/15 10:36, Sudeep Holla wrote: > of_get_cpu_node provides the device node associated with the given > logical CPU and cpu_logical_map contains the physical id for each CPU > in the logical ordering. Passing cpu_logical_map(cpu) to of_get_cpu_node > is incorrect. > > This patch fixes the issue by passing the logical CPU number to > of_get_cpu_node > > Cc: Florian Fainelli <f.fainelli@gmail.com> > Cc: bcm-kernel-feedback-list@broadcom.com > Signed-off-by: Sudeep Holla <sudeep.holla@arm.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Acked-by: Florian Fainelli <f.fainelli@gmail.com> > --- > arch/arm/mach-bcm/bcm63xx_smp.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > Hi, > > I found this accidentally when searching for cpu_logical_map. It's > clearly a bug, it seem to be working fine as the physical and the > logical CPU ordering is same most of the time. Good catch! Kevin, Arnd, Olof, if you have any "late late" fixes for 4.2-rcX, can you pick this one directly? Thanks! > > Regards, > Sudeep > > diff --git a/arch/arm/mach-bcm/bcm63xx_smp.c b/arch/arm/mach-bcm/bcm63xx_smp.c > index 3f014f18cea5..b8e18cc8f237 100644 > --- a/arch/arm/mach-bcm/bcm63xx_smp.c > +++ b/arch/arm/mach-bcm/bcm63xx_smp.c > @@ -127,7 +127,7 @@ static int bcm63138_smp_boot_secondary(unsigned int cpu, > } > > /* Locate the secondary CPU node */ > - dn = of_get_cpu_node(cpu_logical_map(cpu), NULL); > + dn = of_get_cpu_node(cpu, NULL); > if (!dn) { > pr_err("SMP: failed to locate secondary CPU%d node\n", cpu); > ret = -ENODEV; >
Le 07/10/15 10:36, Sudeep Holla a écrit : > of_get_cpu_node provides the device node associated with the given > logical CPU and cpu_logical_map contains the physical id for each CPU > in the logical ordering. Passing cpu_logical_map(cpu) to of_get_cpu_node > is incorrect. > > This patch fixes the issue by passing the logical CPU number to > of_get_cpu_node > > Cc: Florian Fainelli <f.fainelli@gmail.com> > Cc: bcm-kernel-feedback-list@broadcom.com > Signed-off-by: Sudeep Holla <sudeep.holla@arm.com> Applied to soc/next, with the proper Fixes tag for this, thanks!
diff --git a/arch/arm/mach-bcm/bcm63xx_smp.c b/arch/arm/mach-bcm/bcm63xx_smp.c index 3f014f18cea5..b8e18cc8f237 100644 --- a/arch/arm/mach-bcm/bcm63xx_smp.c +++ b/arch/arm/mach-bcm/bcm63xx_smp.c @@ -127,7 +127,7 @@ static int bcm63138_smp_boot_secondary(unsigned int cpu, } /* Locate the secondary CPU node */ - dn = of_get_cpu_node(cpu_logical_map(cpu), NULL); + dn = of_get_cpu_node(cpu, NULL); if (!dn) { pr_err("SMP: failed to locate secondary CPU%d node\n", cpu); ret = -ENODEV;
of_get_cpu_node provides the device node associated with the given logical CPU and cpu_logical_map contains the physical id for each CPU in the logical ordering. Passing cpu_logical_map(cpu) to of_get_cpu_node is incorrect. This patch fixes the issue by passing the logical CPU number to of_get_cpu_node Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: bcm-kernel-feedback-list@broadcom.com Signed-off-by: Sudeep Holla <sudeep.holla@arm.com> --- arch/arm/mach-bcm/bcm63xx_smp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) Hi, I found this accidentally when searching for cpu_logical_map. It's clearly a bug, it seem to be working fine as the physical and the logical CPU ordering is same most of the time. Regards, Sudeep